community pool / model answer

Qwen3-8B

chat · Qwen/Qwen3-8B

52.3 tok/s · model median · 44 runs in the model record.

Reference rig	Quantization	Median result	Basis	TTFT (ms)	Peak VRAM	Max context	Engines
AMD Ryzen 7 7840HS	4-bit	12.3 tok/s	reported	-	- GB	512	llama.cpp
GTX 1080 Ti 11GB	4-bit	50.7 tok/s	reported	-	7.8 GB	8,192	llama.cpp
GTX 1080 Ti 11GB	5-bit	36.8 tok/s	reported	-	8.5 GB	8,192	llama.cpp
Multi-GPU 24GB ×2	4-bit	113.5 tok/s	reported	186	- GB	2,048	llama.cpp
Radeon 890M 32GB	4-bit	15.9 tok/s	reported	-	- GB	4,096	llama.cpp
RTX 3060 12GB	4-bit	43.7 tok/s	measured	370	- GB	131,072	llama.cpp
RTX 3060 Ti 8GB	4-bit	53 tok/s	measured	1,206	7.5 GB	65,536	llama.cpp
RTX 3090 Ti 24GB	4-bit	133 tok/s	reported	123	7 GB	8,192	llama.cpp
RTX 5090 32GB	4-bit	227.1 tok/s	reported	18	5.6 GB	8,192	llama.cpp

Context tested

The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.